14 Replies Latest reply on Aug 6, 2010 10:23 AM by Ryan Werner

    Problem with beam to solid body analysis

    Philip Lewis

      I am pretty new to simulation and need some help. I have what I think is a pretty simple weldment base. All one level with just 2 plates.What keeps happening is when I run the study the plates (where the load is being applied) keep going through the beams and I get huge numbers on the stress values. I have the bottom face of the plate bonded to the beams that they are touching and I am using the outer most nodes as the fixtures as a worst case. I have tried using a mid-surface and bonding to it but the results are similar. Am I doing something wrong? Is there a better way to do this?

        • Re: Problem with beam to solid body analysis
          Devi Prasad Samal

          Hi Phillip,

           

          Whatever u write from that i guess u dont have done anything wrong. but still from your result scenario it seems that some contacts have not been proper.

           

          just some question which arises: how u define bond between the beam and shell elements. which version of solidworks u are using?

           

          and can u post the files with simulation so that we can check it out?

           

          Regards

          Devi Prasad Samal

            • Re: Problem with beam to solid body analysis
              Philip Lewis

              Hi Devi

               

              I have attached the 2 studies I have run. Study 1 has the extruded plate bonded to the beam elements it is in contact with. Study 2 has the surface bonded to the beam in the same manner. I'm using SW2010 SP3.0.

               

              If you or anybody could shed some light on what's going on it would be greatly appreciated.

                • Re: Problem with beam to solid body analysis
                  Devi Prasad Samal

                  hey phillip u have not uploaded the assembly file..

                   

                  with out solidworks file it is difficult to find out something with only simulation result files.

                    • Re: Problem with beam to solid body analysis
                      Philip Lewis

                      Sorry about that. I thought I had the base weldment in the zip file. I have updated the zip file to include the weldment part file.

                        • Re: Problem with beam to solid body analysis
                          Devi Prasad Samal

                          sorry phillip,

                           

                          i have tried to download files but dont know why ur files downloading stop in between middle. so could not have see ur files. i think having bigger size gives me problem. having some restriction on my company internet.

                           

                          please just attach only solidworks files. apply the boundary condition but did not include result files. i just want to check ur simulation setup.

                           

                          sorry again for kep u waiting & not having a reply...

                           

                          thanks...

                            • Re: Problem with beam to solid body analysis
                              Philip Lewis

                              Here is just the solidworks file. Sorry about the download issues.

                                • Re: Problem with beam to solid body analysis
                                  Ryan Werner

                                  Hi Philip,

                                   

                                  I hope you do not mind my looking at this problem.  I have had similar issues trying to run mixed meshes like this in the past.  I believe your problem is not in how you set anything up but instead it has to do with some of the limitations Simulation still has with creating bonds between solids and beams.  I have been in contact with a SW representative about this before and he has told me that what you are trying to do may not work depending on how much surface area you have between the beam and the solid.  Sure enough, I increased the width of your plates to 16" and all the contacts worked fine.  Of course this does not help you at all but at least you know that you had the correct methods.

                                   

                                  My personal recommendation is to try and run the entire model as a solid.  This will take more resources so you may not be able to do it.  But if possible i would give this a try.  You may also run into problems with stress concentrations showing very high stresses but if all you are looking for is a general stress overview then this will give you a good idea of what kinds of stress are present.

                                   

                                  Ryan W.

                                    • Re: Problem with beam to solid body analysis
                                      Philip Lewis

                                      Thanks for looking at it. Thats why I posted here.

                                       

                                      I think you may be right. I heard back from my VAR and he said it worked as a solid body as well. When I increased the width of the plates and ran it as a beam it did make a difference but I still am not sure that it is working correctly. He is sending it in to solidworks. I don't understand how there would be such an issue given that both Cosmos/Simulation and weldments have been around for years.

                                        • Re: Problem with beam to solid body analysis
                                          Ryan Werner

                                          It is a little frustrating to be given the functionality in Simulation to run mixed solids and beams analysis but then be limited in what you can actually get relevant results for.  From what our VAR tells me it sounds like it is still very much a work in progress.  I have more or less taken the view that if possible I run my weldments as solids.  I have found it a whole lot faster and more reliable than trying to get the mixed analysis to run.  It is not always the perfect solution but at least when I get results I can be fairly confident that they are accurate and realistic.

                                            • Re: Problem with beam to solid body analysis
                                              Philip Lewis

                                              It's funny because this is just one of three bases we are using for the same system layout and it is the only one that gave me this problem.  The first one had a beam on a beam which gave me an issue but changing just those elements to solids seemed to solve the problem. The most complicated base didn't give me any trouble at all. It is really kind of strange what works and what doesn't.

                                                • Re: Problem with beam to solid body analysis
                                                  Ryan Werner

                                                  The last base is much more complex.  You would think that of any it would have been the most difficult to get results for.

                                                   

                                                  As far as those beams on top of beams, there is a way to get those to work without having to turn them to solids.  You may already know this but you can actually manually manipulate which members are included in a joint.  You just go into the Edit Joints window and right click on one of the "pink" joint balls.  The included members will highlight and a small window will appear with their names.  You can then add or delete from that joint at will or completely delete the joint altogether.  The tricky part about this technique is that once you are satisfied with your joint you must hit the Calculate button for the changes to stick.  Unfortunately, hitting the Calculate button will calculate new joints for the entire model so if you are doing this for multiple joints you must re-do each joint everytime you make a change or if the joints get out of date.  Needless to say this becomes very tedious but in the end is actually a very powerful tool that SolidWorks doesn't really tell you exists.  I wouldn't know about it unless I hadn't specifically asked my VAR how they could possiblly not have the option.

                                                   

                                                  In your instance you would have to set the Clearance small enough that end joints ("gold/yellow" balls) would be created at the end of the short I beams and then you could go in and edit the joint to include the larger beam underneath it.  This creates very short "fake" beams between the two but does connect them.  Of course it does not truly represent the connection in real life but it does allow forces or restraints to be transmitted through them and when viewing the stress results the beams are there and not missing as they are when they are solids.  

                                                    • Re: Problem with beam to solid body analysis
                                                      Philip Lewis

                                                      I had seen the joints in the simulation training manual but thought that these points were too far from the end of the lower beams. I guess I just didn't realize it could be done at that location. My solid bodies sim is still running so I think I'll try that tomorrow. I guess that's what happens when you go through a 500 page training manual in just 2 1/2 days.

                                                      • Re: Problem with beam to solid body analysis
                                                        Philip Lewis

                                                        Here is what my VAR got back from solidworks:

                                                         

                                                        " Your SR 1-1735124561 (VAR SR# ) has been updated.

                                                        For details, review under the Comments area of My SR's (Service Request) on the portal. contacts between shell and beams appears to behave incorrectly

                                                         

                                                        Hi Doug,

                                                         

                                                        If I mesh with a slightly coarser mesh, the bonding does work. This is most likely because of the larger tolerance of the shell elements. The geometry is such that the shells do not project onto the neutral axis - the edges of the shell are offset to the side of the neutral axis of the beams.

                                                         

                                                        As I mentioned over the phone, we recommend modeling the beams as either solids or shells in such cases in order to get an accurate representation of the bonding to the outer portion of the beam flange face.

                                                         

                                                        I hope this is helpful."

                                                         

                                                        I must say that seems like a really weak reason on their part. Solidworks has had both of these (Cosmos/Simulation and Weldments) running for years and they haven't been able to figure out a way to make this work without another work around.

                                                          • Re: Problem with beam to solid body analysis
                                                            Ryan Werner

                                                            Sounds like a standard VAR reply.  There is always a work around of course but usually I have a pretty good reason why I wanted to run the analysis the way I did in the first place and it is annoying to have to go some other direction because of a glitch.  Oh well, it seems like no matter what you are trying to do in FEA you are always having to bend you wishes to the programs limitations.

                                                             

                                                            Ryan W.